System and method for implementing a trusted dynamic launch and trusted platform module (TPM) using secure enclaves

BACKGROUND

p-00021. Field of the Invention

p-0003This invention relates generally to the field of computer systems. More particularly, the invention relates to a system and method for implementing a trusted dynamic launch and trusted platform module (TPM) using secure enclaves.

p-00042. Description of the Related Art

p-0005The assignee of the present patent application has designed a trusted computing platform known as Trusted Execution Technology (“TXT”) within some microprocessors and respective chipsets to provide computer users and computer system providers with a higher level of trust and control over computer systems. Currently, TXT is employed as a way to defend against software-based attacks aimed at accessing sensitive information. Although often used as a security technology, TXT may also be used to enable development of more advanced, tamper-resistant forms of digital rights management (DRM) and can be used to achieve vendor lock-in for hardware platforms.

p-0006Current TXT implementations consist of a set of hardware enhancements allowing the creation of multiple separated execution environments, sometimes referred to as “partitions.” One particular hardware component used with current TXT implementations is known as the Trusted Platform Module (TPM), which provides secure key generation and storage, as well as authenticated access to data encrypted by generated keys. The private key stored in the TPM is generally not available to the owner of the computer system, and never leaves the TPM chip under normal operation. The TPM that manages Trusted Platform requests generates keys and certificates for private environments (e.g., applications or service spaces) and manages the machine trust state to allow, for example, a local user (or even a remote party) to check the security on a workstation with a higher level of confidence using the Remote Attestation Protocol.

p-0007As illustrated in <figref idrefs="DRAWINGS">FIG. 1</figref>, a TPM module <b>100</b> includes an I/O port <b>120</b> for receiving TPM commands from the execution environment, processing the commands and communicating the results over a system bus (typically a low pin count (LPC) bus). An execution engine <b>104</b> executes program code <b>107</b> in response to the TPM commands and utilizes a variety of different TPM components to perform the necessary operations including a secure hash algorithm 1 (SHA-1) component <b>109</b> for performing SHA-1 hash operations; a random number generator <b>110</b> for generating random numbers; an attestation identify key (AIK) component <b>111</b> for securely establishing that a remote entity is communicating with the TPM; an RSA engine <b>105</b> for implementing RSA encryption and digital signatures; and a key generation module <b>106</b> for generating keys.

p-0008The TPM is an “opt-in” device meaning that the platform owner must take specific steps to turn the TPM on. An opt-in module <b>108</b> securely stores the platform owner's selection of the state of the TPM.

p-0009In addition, a non-volatile memory <b>101</b> and a volatile memory <b>102</b> are provided to store long term and temporary values, respectively, while executing TPM commands. Finally, a set of platform control registers (PCRs) are provided to keep track of measurements reported to the TPM. The <i>PC Client Spec </i>(TCG 2005a) specification requires a minimum of 24 PCRs. Reading and writing to a PCR requires specific TPM ordinals. When an entity needs to store a measurement in the TPM, the TPM provides an assurance that no other entity can change the measured value by not allowing any entity to write directly to the TPM. Rather, the entity “extends” the specified PCR. The extend operation concatenates the current PCR value with the new value and performs an SHA-1 hash on the concatenated value. The resulting hash value is the new PC value. A detailed description of the TPM can be found, for example, in David Grawrock, <i>The Intel Safer Computing Initiative, Building Blocks for Trusted Computing</i>, Intel Press (2006) (hereinafter “Grawrock”) at 119-142, and in the <i>PC Client Spec </i>referenced above.

p-0010For effective security it is important to know what causes the launch of a protected partition. Consequently, the CPU protection modes must receive proper initialization. To this end, current TXT implementations use Safer Mode Extensions (SMX) and Virtual Mode Extensions (VMX) and controls them through the CPU commands GETSEC [SENTER] and GETSEC [SEXIT]. A detailed description of these commands can be found in Grawrock at pages 185-212. Briefly, GETSEC is the CPU instruction that implements SMX extensions. The GETSEC [SENTER] operation provides for the creation of a secure partition during runtime, without the need to reboot the computer system. The GETSEC [SEXIT] instruction is then used to ensure the complete removal of all CPU state associated with the protected partition.

p-0011Currently, a dedicated TPM chip is required to implement a TXT environment as described above. As such, TXT implementations are not typically used on small form-factor, low cost computing platforms. Thus, to reduce the cost and complexity associated with TXT implementations, it would be beneficial to provide a solution which does not require a dedicated TPM chip.

BRIEF DESCRIPTION OF THE DRAWINGS

p-0012A better understanding of the present invention can be obtained from the following detailed description in conjunction with the following drawings, in which:

p-0013<figref idrefs="DRAWINGS">FIG. 1</figref> illustrates a prior art trusted platform module which may be used within a Trusted Execution Technology (TXT) environment.

p-0014<figref idrefs="DRAWINGS">FIG. 2</figref> illustrates one embodiment of an apparatus for implementing a trusted platform module (TPM) using secure enclaves.

p-0015<figref idrefs="DRAWINGS">FIG. 3</figref> illustrates a secure enclave control store and an enclave memory in accordance with one embodiment of the invention.

p-0016<figref idrefs="DRAWINGS">FIG. 4</figref> illustrate proxy and bridge edge routines implemented in one embodiment of the invention.

p-0017<figref idrefs="DRAWINGS">FIG. 5</figref> illustrates one embodiment of a conceptual layout of processor reserved memory (PRM) and an enclave page cache (EPC).

DETAILED DESCRIPTION

p-0018In the following description, for the purposes of explanation, numerous specific details are set forth in order to provide a thorough understanding of the embodiments of the invention described below. It will be apparent, however, to one skilled in the art that the embodiments of the invention may be practiced without some of these specific details. In other instances, well-known structures and devices are shown in block diagram form to avoid obscuring the underlying principles of the embodiments of the invention.

p-0019One embodiment of the invention provides a software-based solution using limited hardware for implementing trusted computing base (TCB) measurement and attestation features. Specifically, one embodiment of the invention uses Secure Enclaves, hardware registers which behave like TPM platform configuration registers (PCRs), and a small amount of non-volatile storage to implement a software-based TPM and provide functionality which is equivalent to the GETSEC [SENTER] instruction. The embodiments discussed below enable the measurement and launch of a Measured Launched Environment and SMI Transfer Monitor (STM). Embodiments of the invention also provide TPM-equivalent functionality by exposing hardware registers and non-volatile RAM to a Secure Enclave-based TPM implementation.

p-0020Briefly, embodiments of the invention provide the above functionality using the following components:

p-0021Platform Configuration Registers (PCRs) provided by hardware primitives which extend and reset in the same manner as TPM PCRs.

p-0022Non-volatile RAM: One embodiment of the invention exposes and controls access to a relatively small amount of non-volatile memory which provides the basis for securing TPM NV-RAM and TXT Launch Control Policy data. The NV-RAM also provides the bases for the TPM monotonic counter features which, as described below, are exposed by the TPM secure enclave.

p-0023TPM Root Key: In order to protect user data secrets and TPM data assets, one embodiment of the invention of the invention provides TPM secure execution with a root key that has the following unique characteristics: (1) the key is platform specific; (2) the key is platform owner specific; and (3) the key is specific to the TPM secure execution environment.

p-0024TPM Secure Execution Environment: One embodiment of the invention uses Secure Enclaves or a similar technology to provide a secure execution environment. This environment hosts a software implementation of the TPM specification using the TPM Key (mentioned above) to secure its data assets, and those of the user. The hardware primitives (also mentioned above) are also exposed securely to the secure execution environment for use by the TPM software.

p-0025Dynamic Root of Trust Measurement (DRTM) Launch Enclave: Embodiments of the invention discussed below also use secure enclaves (or similar technology) to provide a software implementation of the TXT GETSEC [SENTER] instruction. This enclave may provide the following features: <ul><li id="ul0001-0001" num="0000"><ul><li id="ul0002-0001" num="0025">1) The DRTM enclave rendezvous all hardware threads on the platform in preparation for the launch of the Measured Launched Environment (MLE).</li><li id="ul0002-0002" num="0026">2) The DRTM enclave prevents errant threads and no-host processing agents from affecting launch</li><li id="ul0002-0003" num="0027">3) The DRTM enclave ensures that Memory Configuration of the platform has been established correctly.</li><li id="ul0002-0004" num="0028">4) The DRTM enclave reports (using the hardware configuration registers) the Launch Policy (i.e. Intel TXT Launch Control Policy) that was applied at launch and which entity (i.e., which DRTM enclave) applied it.</li><li id="ul0002-0005" num="0029">5) The DRTM enclave verifies and sets the platform hardware configuration required to protect the MLE (e.g., the Intel Virtualization Technology for Directed I/O (VT-d) tables for DMA protection). Currently, VT-d enables protection by restricting direct memory access (DMA) of devices to pre-assigned domains or physical memory regions. This is achieved by a hardware capability known as DMA-remapping.</li><li id="ul0002-0006" num="0030">6) The DRTM enclave computes a cryptographic measurement of the system management interrupt (SMI) transfer monitor, and records the measurement using the hardware configuration registers.</li><li id="ul0002-0007" num="0031">7) The DRTM enclave computes a cryptographic measurement of the MLE, records that measurement using the hardware configuration registers and passes execution control to the MLE.</li></ul></li></ul>

p-0026One embodiment of a system architecture for implementing a software-based TPM <b>212</b> within a secure enclave <b>250</b> is illustrated in <figref idrefs="DRAWINGS">FIG. 2</figref>. As mentioned above, the secure enclave <b>250</b> is a trusted software execution environment which prevents software executing outside the enclave from having access to software and data inside the enclave. One embodiment of a secure enclave is described in the co-pending U.S. patent application entitled Method and Apparatus to Provide Secure Application Execution, filed Nov. 13, 2009, Ser. No. 12/590,767, and in the co-pending PCT Patent Application entitled Method and Apparatus to Provide Secure Application Execution, filed Dec. 22, 2009, Application No. PCT/US2009/069212, which are both incorporated herein by reference. These applications are referred to herein collectively as the “co-pending applications.”

p-0027This application will now describe certain aspects of secure enclave design as described in the co-pending applications and illustrated in <figref idrefs="DRAWINGS">FIGS. 3-5</figref>. Following this description, a detailed description of implementing a TPM using a secure enclave will be provided with respect to <figref idrefs="DRAWINGS">FIG. 2</figref>.

p-0028As described in the co-pending applications, a secure enclave <b>250</b> is carved out of the virtual address space of its containing process. A complete description of secure enclave creation, management and termination can be found in the co-pending applications. All memory management of the enclave is performed in terms of virtual addresses. Enclave memory management is divided into two parts: address space allocation and memory commitment. Address space allocation is the specification of a maximal range of addresses that the enclave may use. No actual resources are committed to this region. Memory commitment is the assignment of actual memory resources (as pages) within the allocated address space. This two-phase technique allows enclaves to flexibly control their memory usage and allow for growth without overusing memory resources when enclave needs are low. Commitment adds virtual pages to the enclave.

p-0029This distinction is reflected in the two instructions ECREATE and EADD. ECREATE allocates a range of addresses for use by the enclave. EADD commits virtual memory pages within that range to the enclave. An operating system may support separate allocate and commit operations. For example, the Windows API VirtualAlloc takes a flag argument of MEM_RESERVE (allocation) or MEM_COMMIT (commitment). It is not required that an OS support this separation, but its presence can make memory management within an enclave more flexible and efficient.

p-0030In one embodiment, a secure enclave <b>250</b> is created using the ECREATE instruction, specifying a Base and size of the virtual address space range in which the enclave is to be built. Memory is added to the secure enclave before it is sealed. One embodiment of enclave memory layout is illustrated in <figref idrefs="DRAWINGS">FIG. 3</figref>.

p-0031The enclave creation and commitment process is best illustrated by example using the Windows API. An address space region is reserved in the process using VirtualAlloc, passing in the MEM_RESERVE flag. This reserves a region of memory <b>301</b> without actually committing any physical memory or page file storage. <ul><li id="ul0003-0001" num="0000"><ul><li id="ul0004-0001" num="0038">1) Commit the first page of the enclave by calling VirtualAlloc again, this time with the MEM_COMMIT flag.</li><li id="ul0004-0002" num="0039">2) Use the ECREATE instruction to set up the initial environment, specifying the same address range as in step 1.</li><li id="ul0004-0003" num="0040">3) For each additional page to be added to the enclave:</li><li id="ul0004-0004" num="0041">4) VirtualAlloc the enclave page <b>301</b>.</li><li id="ul0004-0005" num="0042">5) Use EADD to add the new page to the enclave.</li><li id="ul0004-0006" num="0043">6) Use EEXTEND to add a measurement for 128 bytes of the page. Call this instruction until the entire enclave is measured.</li><li id="ul0004-0007" num="0044">7) If a contiguous set of enclave pages is to be added at once, the above steps can be reordered and optimized to minimize the number of system calls.</li><li id="ul0004-0008" num="0045">8) On operating systems that do not support separate reserve and commit operations, the VirtualAlloc in step 1 can be replaced by a simple malloc, for example, and the remaining calls to VirtualAlloc eliminated.</li></ul></li></ul>

p-0032Because the ECREATE and EADD instructions enable encryption and integrity for the added page, it is not possible for non-enclave software to initialize this memory after it has been added to the enclave. The runtime system must completely initialize any memory for which it is responsible before EADDing it. This typically requires the runtime system to build the enclave image in-place, even before calling ECREATE (i.e., because SECS cannot be modified by software after ECREATE). Once memory has been initialized, the enclave ECREATEd and its memory EADDed, it may be sealed and measured as described herein.

p-0033In one embodiment, there are two data structures inside the enclave, the Secure Enclave Control Store (SECS) <b>302</b> and the Thread Control Structure (TCS) <b>304</b>. SECS and TCS contain architectural information and non-architectural convenience fields used by the runtime software. All of the convenience fields and a few of the architectural fields are initialized by software. During enclave creation, code <b>306</b> and data <b>305</b> for an enclave may be loaded from a clear-text source. It is not possible to load encrypted code and/or data directly into an enclave during creation. However, code and/or data from an encrypted source may be installed into an enclave by first loading a trusted loader into the enclave. Once the enclave is running, this loader can then be used to install secret code/data into the enclave. Details of this procedure are beyond the scope of the present application.

p-0034Once the SECS <b>302</b>, one or more TCSs <b>304</b> and the code <b>306</b> and data <b>305</b> have been EADDed, the enclave creation is completed by the EINIT instruction. This instruction stops further measurement information from being written into a measurement register (MR) (sometimes referred to in the co-pending applications as “IR0”). After EINIT no further pages may be added to the enclave.

p-0035Untrusted code calls into the enclave using the EENTER instruction and the enclave returns back to the untrusted caller via EEXIT. On enclave entry, control is transferred to code which switches the stack pointer to one inside the enclave. When returning the software again swaps the stack pointer then executes the EEXIT instruction.

p-0036In non-enclave mode, subroutine parameters are generally pushed onto the stack. The called routine, being aware of its own stack layout, knows how to find parameters based on compile-time-computable offsets from the SP or BP register (depending on runtime conventions used by the compiler).

p-0037Because of the stack switch when calling an enclave, stack-located parameters cannot be found in this manner. Edge routines (untrusted procedures that ECALL and trusted procedures that are the target of ECALL) are required to define a different parameter passing convention for parameters that cannot fit in registers. For example, the caller might push parameters onto the untrusted stack and then pass a pointer to those parameters in AX. The exact choice of calling conventions is up to the writer of the edge routines; be those routines hand-coded or compiler generated. As illustrated in <figref idrefs="DRAWINGS">FIG. 4</figref>, the proxy routine <b>401</b> (on the caller's side of the enclave boundary) is responsible for marshalling parameters in a runtime-system-defined format and passing them to the bridge routine <b>402</b> on the other side of the boundary which hands off to a trusted function <b>403</b>.

p-0038As with most systems, it is the responsibility of the callee to preserve all registers except that used for returning a value. This is consistent with conventional usage and tends to optimize the number of register save/restore operations that need be performed. It has the additional security result that it ensures that data is scrubbed from any registers that were used to temporarily contain secrets.

p-0039No registers are modified during EEXIT. This requires that the software scrub all secrets from registers before returning. This could be done in the bridge function <b>402</b> or it could simply rely on the standard register preservation across procedure calls that compilers typically emit as prolog/epilog.

p-0040Calling an untrusted procedure from an enclave also uses the EEXIT instruction. While there is a stack switch back to the untrusted stack, the EEXIT contains a destination address which is the procedure.

p-0041In order to avoid leaking secrets though values left over in registers, the trusted software clears all computation registers except for AX (which may be used for parameter passing). This is necessary because even the most aggressively optimizing compilers do not perform the flow analysis to answer the question—did register X ever contain an intermediate value during the lifetime of this call tree? Thus, it cannot be determined which registers might contain secrets and which are safe. As a result of this clearing, edge routines save registers before the EEXIT and restore them after the EENTER returns. No registers are modified during EENTER. Since there are no secrets generated or used by the untrusted CALLee, this is not necessary.

p-0042To destroy the enclave, the EPC manager, either a driver or the OS can reclaim the enclave's EPC memory by using the EREMOVE instruction. This removes a page of memory from the EPC. The EPC manager uses EREMOVE on every page.

p-0043As illustrated in <figref idrefs="DRAWINGS">FIG. 5</figref>, the Enclave Page Cache (EPC) <b>403</b> is a secure storage used by the CPU to store enclave pages when they are a part of executing enclave. In one embodiment, the following requirements are identified on the EPC. <ul><li id="ul0005-0001" num="0000"><ul><li id="ul0006-0001" num="0058">1. Any accesses to the enclave memory pages loaded into the EPC that belong to non-debug enclaves must be protected from any modification by software entities outside that enclave.</li><li id="ul0006-0002" num="0059">2. Attackers should not be able to read plain-text data belonging to non-debug enclaves that is loaded into the EPC via straight-forward hardware attacks.</li><li id="ul0006-0003" num="0060">3. Attackers should not be able to able to modify data in the EPC that belongs to non-debug enclaves via straight-forward hardware attacks.</li><li id="ul0006-0004" num="0061">4. Any data loaded into the EPC must be accessible coherently, yet securely from any CPU in the system.</li></ul></li></ul>

p-0044There are several mechanisms of implementing the EPC <b>503</b>. The EPC may be implemented as on on-die SRAM or eDRAM. The EPC may also be constructed by dynamically sequestering ways of the CPU's last-level cache. In such an implementation, the EPC must be protected from un-authorized accesses from outside the package. However, other packages in the system must be able to access the EPC coherently, yet securely.

p-0045Another mechanism of implementing EPC <b>503</b> is to implement a Memory Encryption Engine (MEE). MEE provides a cost-effective mechanism of creating cryptographically protected volatile storage using platform DRAM. MEE uses one or more strategically placed cryptographic units in the CPU uncore to provide varying levels of protection, as needed by the customer technology. The various uncore agents are modified to recognize the memory accesses going to the MEE, and to route those accesses to a Crypto Controller located in the uncore. The Crypto Controller, depending on the desired protection level, generates one or more memory accesses to the platform DRAM to fetch the cipher-text. It then processes the cipher-text to generate the plain-text, and satisfies the original MEE memory request. MEE fully integrates into the Intel QuickPath Interconnect (QPI) protocol, and scales to multi-package platforms, with security extensions to the QPI protocol. In a multi-package platform configuration, MEE protects memory transfers between Intel CPUs using a crypto engine in the externally facing QPI link layers.

p-0046In one embodiment, the EPC is divided into chunks of 4 KB pages and is aligned into a 4 KB boundary. Each 4 KB chunk of EPC is called an EPC page. EPC is used to hold pages belonging to each running instance of an enclave. Pages in the EPC can either be valid or invalid. Every valid page in the EPC belongs to exactly one running instance of an enclave. Each running instance of an enclave in turn has exactly one EPC page holding its SECS <b>302</b>.

p-0047The security attributes for each EPC page are held in an internal structure called Enclave Page Cache Map (EPCM) which is a secure structure used by the processor to track the contents of the EPC. The EPCM holds exactly one entry for each page that is currently loaded into the EPC. The format of the EPCM is microarchitectural, and consequently is implementation dependent. However, architecturally, the EPCM holds following information: <ul><li id="ul0007-0001" num="0000"><ul><li id="ul0008-0001" num="0066">1. An indication whether the corresponding EPC page is valid.</li><li id="ul0008-0002" num="0067">2. If the page is valid, which running instance of enclave it belongs to. This is essentially back pointer to the SECS of the running instance of the enclave. A number of SE instructions follow this back pointer to read the contents of the SECS page.</li><li id="ul0008-0003" num="0068">3. If the page is valid, the type of the corresponding EPC page. Three types are supported: Regular Page (PT_REG), TCS Page (PT_TCS), and SECS Page (PT_SECS).</li><li id="ul0008-0004" num="0069">4. If the page is valid, the linear address through which the enclave is allowed to access the page.</li><li id="ul0008-0005" num="0070">5. If the page is valid, the enclave-specified read/write/execute restrictions on that page, if any.</li></ul></li></ul>

p-0048The EPCM structure is used by the CPU in the address-translation flow to enforce access-control on the enclave pages loaded into the EPC. The EPCM structure is described in the co-pending applications, as is the conceptual access-control flow. The EPCM entries are managed by the processor as part of various instruction flows.

p-0049EPC <b>502</b> layout is specific to a model, and can be enumerated through various MSR. The co-pending applications describe enumeration for the mechanism of enumerating the EPC layout. At a high level, a CPU that supports SE supports the ability for the BIOS to reserve a physically contiguous range of memory called Processor Reserved Memory (PRM). PRM must have a size that is integer power of two, and must be naturally aligned. The BIOS identifies the PRM by setting a pair of MSRs, collectively known as the PRMRR <b>501</b>.

p-0050Within the PRM, the processor supports one or more sections of the Enclave Page Cache (EPC). Each section of enclave page cache starts on a 4 KB boundary, and has a size that is integer multiple of 4 KB. Each section of EPC also has a security level associated with it. In one embodiment, two Security Levels are defined: Confidentiality-protected EPC (SL_CP), and Replay protected EPC (SL_RP). <figref idrefs="DRAWINGS">FIG. 5</figref> illustrates one embodiment of a memory layout.

p-0051Note that some of the details from <figref idrefs="DRAWINGS">FIGS. 3-5</figref> are not shown in <figref idrefs="DRAWINGS">FIG. 2</figref> to avoid obscuring the underlying principles of the invention. In contrast to the GETSEC [SENTER] instruction which is capable of launching one measured launched environment at a time, multiple secure enclaves <b>250</b> may be executed within a single OS environment <b>251</b> (although only a single enclave is shown in <figref idrefs="DRAWINGS">FIG. 2</figref> for simplicity). In addition, the mechanism for delivering a key into the enclave is inherent to the enclave, whereas with existing TXT implementations, the TPM creates a key and only releases it if a dynamic launch event has occurred.

p-0052As discussed in detail below, in one embodiment, native hardware capabilities <b>290</b> are provided to build and manage the secure enclaves <b>250</b>. The term “native” in this context means that hardware capabilities are provided by a computer processor and/or chipset component (i.e., a separate, dedicated TPM chip is not required for implementing this embodiment of the invention). As shown in <figref idrefs="DRAWINGS">FIG. 2</figref>, the hardware capabilities include a set of platform PCR (pPCR) registers <b>201</b>, Flags <b>202</b> (sometimes referred to as Intel Architecture (IA) Flags), a platform unique key <b>203</b>, an owner epoch <b>204</b>, attestation primitives <b>291</b> and a monotonic counter <b>207</b>.

p-0053In a TPM device, as the computing platform boots, a hash is written for the data and program code which the TPM is about to process into a PCR register in the TPM. Consequently, in order to implement a software-based TPM <b>212</b>, a measurement is initially sent to the pPCRs <b>201</b> in the early boot stages of the computing system. As is known by those of skill in the art, the PCR registers <b>201</b> are capable of performing “extend” operations while writing data. Specifically, an extend operation takes a value to be written and the previously stored value cryptographically hashes then together. As shown in <figref idrefs="DRAWINGS">FIG. 2</figref>, the set of pPCR <b>201</b> registers are provided within the native hardware based capabilities <b>290</b> to provide the same (or similar) functionality as standard TPM PCR registers. As the computing platform boots, it writes to the pPCR registers which performs extend operations while writing the data. Once the TPM <b>212</b> within the secure enclave <b>250</b> is up and running, it reads the pPCR register values into a storage area in memory, shown in <figref idrefs="DRAWINGS">FIG. 2</figref> as temporary PCRs (tPCRs) <b>215</b>.

p-0054In one embodiment, the operating system (OS) environment <b>251</b> and applications <b>220</b> which are hosted on the computing platform are untrusted by the secure enclave <b>250</b>. Consequently, the secure enclave <b>250</b> needs to have a trusted path to read in data (e.g., data from the pPCRs). One embodiment of the invention uses the EGETKEY command described in the co-pending applications for reading hardware secrets. In particular, the EGETKEY command allows the secure enclave <b>250</b> to bring data directly into the enclave without an external OS call, as illustrated by the trusted paths <b>280</b>-<b>281</b> in <figref idrefs="DRAWINGS">FIG. 2</figref>. By way of example, and not limitation, the EGETKEY command may be used by the secure enclave <b>250</b> to read in data from the pPCRs into the tPCRs after the secure enclave <b>250</b> has been initialized. In one embodiment, the pPCR registers are registers within a central processing unit dedicated to performing PCR functions and the tPCRs are a software implementation of the pPCR registers used by the software-based TPM device <b>212</b>.

p-0055In one embodiment, the secure enclave root key <b>218</b> (sometimes referred to as the “seal key” in the co-pending applications) is generated using data from Flags <b>202</b>, the platform unique key <b>203</b>, the owner epoch value <b>204</b>, and attestation primitives <b>291</b>. In one embodiment, the EGETKEY command combines these values together using an Advanced Encryption Standard (AES)-CMAC function (which is an National Institute of Standards and Technology (NIST) approved algorithm for deriving multiple keys from a single unique value).

p-0056The Flags <b>202</b> shown in <figref idrefs="DRAWINGS">FIG. 2</figref> are flags associated with the secure enclave <b>250</b> indicating a particular mode of operation such as debug mode and/or an indication that the secure enclave <b>250</b> has been initialized. These are sometimes referred to as “attributes” of the enclave as defined in the co-pending applications. The Flags may be stored in a memory such as a register on the CPU.

p-0057In one embodiment, each computing platform is provided with a platform unique key <b>203</b> which may be permanently stored in the CPU and/or chipset components (e.g., by programming fuses). Enclave-specific keys such as the root key <b>218</b> are derived (in part) from the platform unique key <b>203</b>. In one embodiment, a secure enclave <b>250</b> uses the EGETKEY command described in the co-pending applications to retrieve the platform unique key <b>203</b>.

p-0058In one embodiment, the owner epoch value <b>204</b> is set in the computer system's basic input output system (BIOS). The user may change the owner epoch value, for example, when transferring the computer to another user. Changing the owner epoch effectively wipes away all of the original user's secrets (i.e., because it results in a new root key <b>218</b>). Without this ability, a new user could gain access to the original user's secrets.

p-0059The attestation primitives <b>291</b> include measurement registers <b>205</b> and a platform attestation key <b>206</b>. As described in the co-pending applications, in one embodiment, two measurement registers <b>205</b> are provided by the secure enclave: MREADD, which is a measurement of the data contained in the secure enclave; and MRPOLICY, which provides a measurement of a public key that signed the measurement of the enclave (i.e., identifying who produced the enclave). The platform attestation key <b>206</b> allows the enclave to demonstrate that the enclave actually exists. Specifically, the platform attestation key <b>206</b> may be used to sign the MREADD and the MRPOLICY values contained in the measurement registers <b>205</b>, thereby demonstrating that the enclave was built on the computing platform. Thus, the platform attestation key performs a similar function to the attestation identify key (AIK) on a TPM device (i.e., the key signs to demonstrate that the TPM has the designated values).

p-0060As described in the co-pending applications, in one embodiment, four commands are used to build the secure enclave <b>250</b>: ECREATE, EADD, EEXTEND and EINIT. Until these four commands are successfully executed, program code cannot be executed within the secure enclave <b>250</b>. The first three (ECREATE, EADD, EEXTEND) build up a cryptographic measurement of the enclave itself. This is done in the memory region (e.g., a page of data) referred to above as the secure enclave control structure (SECS), which is specific to each enclave. As pages are built into the enclave the measurement registers are updated to include a cryptographic measurement of all of the data inside the enclave. Thus, by the time the EINIT command is executed, there is an identify value of everything stored in the enclave. Like the PCRs, the measurement registers <b>205</b> extend using hash values.

p-0061When all of the enclave pages have been added to the enclave, the OS environment <b>251</b> will execute the EINIT instruction initializing the enclave. A parameter to the EINIT instruction is a permit which demonstrates that the enclave is licensed to run on that machine. When an application is loaded a permit needs to be created. After EINIT successfully completes, the application can execute the EENTER command to enter the enclave.

p-0062In order to preserve data between instantiations of the secure enclave (e.g., platform resets) the root key <b>218</b> will be the same on each instantiation. A user may then encrypt and store data in external storage <b>210</b> (hard drive) and decrypt the data on a subsequent instantiation.

p-0063One embodiment of the invention includes a monotonic counter <b>207</b> which may be used for digital rights management on the computing platform. Specifically, the monotonic counter <b>207</b> may be used to prevent anti-replay attacks on DRM protected data. For example, if a user is only permitted to watch a particular movie three times, the user may make a copy after the second viewing (or sooner) and then replace it on the hard drive in an attempt to circumvent the DRM algorithm. To prevent this circumvention, as illustrated in <figref idrefs="DRAWINGS">FIG. 2</figref>, in one embodiment, the monotonic counter <b>220</b> may be used to count the number of times a file is accessed (e.g., played back in the case of a video) and determine if the file currently being accessed is the most recent version. In one embodiment, a set of software-based monotonic counters <b>214</b> are implemented within the TPM <b>212</b>. The TPM <b>212</b> virtualizes the software based monotonic counters <b>214</b> out of the hardware-based monotonic counter <b>220</b>.

p-0064As illustrated in <figref idrefs="DRAWINGS">FIG. 2</figref>, the TPM <b>212</b> includes a storage root key <b>217</b> used to encrypt all other keys within the TPM architecture. Also shown in <figref idrefs="DRAWINGS">FIG. 2</figref> is a non-volatile memory <b>213</b> as required for implementing a TPM. Finally, an owner module <b>216</b> is used to identify the TPM's owner.

p-0065Other embodiments of the invention may be implemented on cellular phones and pagers (e.g., in which the software is embedded in a microchip), handheld computing devices (e.g., personal digital assistants, smartphones), and/or touch-tone telephones. It should be noted, however, that the underlying principles of the invention are not limited to any particular type of communication device or communication medium.

p-0066Embodiments of the invention may include various steps, which have been described above. The steps may be embodied in machine-executable instructions which may be used to cause a general-purpose or special-purpose processor to perform the steps. Alternatively, these steps may be performed by specific hardware components that contain hardwired logic for performing the steps, or by any combination of programmed computer components and custom hardware components.

p-0067Elements of the present invention may also be provided as a computer program product which may include a machine-readable medium having stored thereon instructions which may be used to program a computer (or other electronic device) to perform a process. The machine-readable medium may include, but is not limited to, floppy diskettes, optical disks, CD-ROMs, and magneto-optical disks, ROMs, RAMs, EPROMs, EEPROMs, magnet or optical cards, propagation media or other type of media/machine-readable medium suitable for storing electronic instructions. For example, the present invention may be downloaded as a computer program product, wherein the program may be transferred from a remote computer (e.g., a server) to a requesting computer (e.g., a client) by way of data signals embodied in a carrier wave or other propagation medium via a communication link (e.g., a modem or network connection).

p-0068Throughout this detailed description, for the purposes of explanation, numerous specific details were set forth in order to provide a thorough understanding of the present invention. It will be apparent, however, to one skilled in the art that the invention may be practiced without some of these specific details. In certain instances, well known structures and functions were not described in elaborate detail in order to avoid obscuring the subject matter of the present invention. Accordingly, the scope and spirit of the invention should be judged in terms of the claims which follow.